The goal of this application is to develop methods to deconvolute the computational complexity of tissue based molecular signatures. In this proposal we will concentrate on the human breast as a model organ and as a methodological proof-of-principle. However, we expect that the approaches we develop will be widely applicable to all tissue and tumor types. B. SIGNIFICANCE: We discovered that the human breast is composed of 11 cell types and 4 hormonal states (HR0-3). Importantly, there was up to 7-fold survival difference in the outcome of patients with HR0 vs. HR3 breast tumors. Our preliminary work suggests that much of the phenotypic differences between HR groups are epigenetic. Thus, a comprehensive characterization of HR0-3 specific epigenetic signatures may have great clinical significance. C. CHALLENGE: The vast majority of existing high-throughput molecular signatures of normal and malignant human tissues are derived from unfractionated tissue fragments. Thus, the existing data is a composite reflecting a tissue mosaic that is composed of many cell types. The question we are tackling here is how to deconvolute the existing Epigenomics and TCGA tissue-specific signatures into their single cell-type-specific components (HR0-3). This is responsive to RFA-RM-14-001 stated purposes: (1) Analyses that use reference epigenomic data to identify specific features that distinguish cell types and (2) Integrative analyses that combine reference epigenomic maps with other public or investigator-generated data sets. C. HYPOTHESIS: (1) identification of lineage specific breast cell-surface markers would permit the isolation of specific cell lineages, facilitating downstream research; and (2) identification of lineage specific epigenetic markers will facilitate development of computational techniques that are necessary for deconvolution of complex epigenomic signatures. AIM 1: Isolation and profiling of HR0-3 cell lineage from normal and malignant human breast 1 A. We will isolate HR0-3 cell subtypes from normal and malignant breast tissues using intracellular markers ER, AR, VDR. 1 B. We will identify cell surface markers that correspond to HR0-3 cell subtypes in normal breast and breast cancers, permitting isolation of viable ER+, vs. AR+ vs. VDR+ vs. HR0 vs. HR3 cell types. AIM 2. Computational determination of lineage-specific and tumor-specific DNA methylation markers in human breast tissue. 2 A. We will characterize tumor-specific and lineage-specific DNA methylation markers. 2 B. We will validate the lineage and tumor specificity of DNA methylation markers by analyzing methylation data from normal breast tissue samples. 2 C. We will investigate the feasibility of computational determination of lineage-specific and tumor-specific DNA methylation markers by integration with other genomic types.